3D Object Completion via Class-Conditional Generative Adversarial Network

Yu-Chieh Chen, Daniel Stanley Tan, Wen-Huang Cheng, Kai-Lung Hua*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference Article in proceedingAcademicpeer-review


Many robotic tasks require accurate shape models in order to properly grasp or interact with objects. However, it is often the case that sensors produce incomplete 3D models due to several factors such as occlusion or sensor noise. To address this problem, we propose a semi-supervised method that can recover the complete the shape of a broken or incomplete 3D object model. We formulated a hybrid of 3D variational autoencoder (VAE) and generative adversarial network (GAN) to recover the complete voxelized 3D object. Furthermore, we incorporated a separate classifier in the GAN framework, making it a three player game instead of two which helps stabilize the training of the GAN as well as guides the shape completion process to follow the object class labels. Our experiments show that our model produces 3D object reconstructions with high-similarity to the ground truth and outperforms several baselines in both quantitative and qualitative evaluations.
Original languageEnglish
Title of host publicationMultiMedia Modeling
Subtitle of host publication25th International Conference, MMM 2019, Thessaloniki, Greece, January 8–11, 2019, Proceedings, Part II
EditorsBenoit Huet, Ioannis Kompatsiaris, Stefanos Vrochidis, Vasileios Mezaris, Wen-Huang Cheng, Cathal Gurrin
Number of pages13
ISBN (Print)9783030057152, 9783030057169
Publication statusPublished - 11 Dec 2018
Externally publishedYes
Event25th International Conference on MultiMedia Modeling - Thessaloniki, Greece
Duration: 8 Jan 201911 Jan 2019

Publication series

SeriesLecture Notes in Computer Science


Conference25th International Conference on MultiMedia Modeling
Abbreviated titleMMM 2019
Internet address


  • Generative adversarial network
  • Object classification
  • Object reconstruction
  • Shape completion


Dive into the research topics of '3D Object Completion via Class-Conditional Generative Adversarial Network'. Together they form a unique fingerprint.

Cite this